Myself Subhanshu Mohan Gupta
I build the infrastructure nobody notices until it breaks at which point everybody notices it.
The rest of the time I argue with YAML and lose to both of my cats.
- 🥅 The 2026 goal was "contribute more to open source." It is September. There are 91 public repos down there. Calling it
- 🤖 Scroll far enough and this README admits 99.98% of last week's code was written by AI. The other 0.02% was me, fixing it
- 👯 Will happily work with anyone who writes the runbook before the incident
- 🐾 Cat dad to Linux and Xenix who share my birthday and have full root on me :)
My Creations (click to expand)
|
|
|
|
|
|
|
|
Open Source (click to expand)
|
🛡️ Security & Compliance eBPF Security Observability Container escape detection, 6 hrs to 90 s Agentic AI EU Compliance 92% of EU AI Act Article 9 mapped to CI gates DevSecOps on AWS EKS CVE turnaround, 14 days to 28 hrs AI-Powered Container Security Risky workloads caught before prod, not after DevSecOps Bootcamp Harden code, pipelines and infra one task at a time |
🤖 AI & Agents Agentic RAG for DevSecOps Retrieval precision, 0.41 to 0.83 AI Research Assistant 12 min of lookup down to 40 s Agents as State Machines One agent and a state machine beats five in costume Terminal Bench: Hard Tier Frontier models score 31% here, 78% on average Shop-U-Now AI Assistant NDCG@5, 0.46 to 0.71 |
☁️ Cloud & Automation On-Call Incident Triage Page to acknowledgement, 9 min to 47 s Multi-Region Tenant Infra New tenant, new region, 2 days to 17 min Survival-Grade Infra Systems that outlive a regional collapse Bitcoin Regtest Pipeline CI confirms a real transaction in 42 s Automated Rolling Updates 60-node fleet upgrades, zero downtime |
📐 Architecture notes (opinions, with diagrams): Distributed Monolith Tax · Short-Lived OIDC for CI · TEE.fail · EU CRA Countdown · The 50ms Edge AI Lie · Agentic SOC, Own vs Buy · Crypto Inventory
🧪 Earlier things: Team Builder · Sorting Visualizer · Currency Convertor · WebRES · Geolocation Finder · GitLab Migration Toolkit
Github Contributions (click to expand)
Wakatime Data & Productivity (click to expand)
🐱 My GitHub Data
📦 1.1 MB Used in GitHub's Storage
🏆 3,040 Contributions in the Year 2026
💼 Opted to Hire
📜 90 Public Repositories
🔑 27 Private Repositories
I'm a Night 🦉
🌞 Morning 2586 commits ████░░░░░░░░░░░░░░░░░░░░░ 17.96 %
🌆 Daytime 3429 commits ██████░░░░░░░░░░░░░░░░░░░ 23.81 %
🌃 Evening 4148 commits ███████░░░░░░░░░░░░░░░░░░ 28.80 %
🌙 Night 4238 commits ███████░░░░░░░░░░░░░░░░░░ 29.43 %
📅 I'm Most Productive on Saturday
Monday 3380 commits ██████░░░░░░░░░░░░░░░░░░░ 23.47 %
Tuesday 1511 commits ███░░░░░░░░░░░░░░░░░░░░░░ 10.49 %
Wednesday 997 commits ██░░░░░░░░░░░░░░░░░░░░░░░ 06.92 %
Thursday 1168 commits ██░░░░░░░░░░░░░░░░░░░░░░░ 08.11 %
Friday 1983 commits ███░░░░░░░░░░░░░░░░░░░░░░ 13.77 %
Saturday 3751 commits ███████░░░░░░░░░░░░░░░░░░ 26.05 %
Sunday 1611 commits ███░░░░░░░░░░░░░░░░░░░░░░ 11.19 %
📊 This Week I Spent My Time On
🕑︎ Time Zone: Asia/Kolkata
💬 Programming Languages:
Other 47 hrs 35 mins ███████████████████████░░ 90.52 %
Markdown 2 hrs 18 mins █░░░░░░░░░░░░░░░░░░░░░░░░ 04.37 %
Python 1 hr 46 mins █░░░░░░░░░░░░░░░░░░░░░░░░ 03.38 %
Text 20 mins ░░░░░░░░░░░░░░░░░░░░░░░░░ 00.65 %
XML 8 mins ░░░░░░░░░░░░░░░░░░░░░░░░░ 00.26 %
🔥 Editors:
Claude Code 49 hrs 18 mins ███████████████████████░░ 93.78 %
VS Code 3 hrs 16 mins ██░░░░░░░░░░░░░░░░░░░░░░░ 06.22 %
💻 Operating System:
Mac 52 hrs 35 mins █████████████████████████ 100.00 %
🤖 AI Coding This Week
⏱ AI Coding Time: 52 hrs 32 mins (99.9%)
✍️ 19,986 lines written by AI, 26 lines written by hand (99.87% AI-written)
🔤 102,689,767 Input Tokens, 9,740,124 Output Tokens
💵 $1039.18 Estimated AI Cost This Week
🧠 3652 AI Sessions, 11651 AI Prompts
Opus 16,516 lines ████████████████████░░░░░ 81.85 %
Fable 2,703 lines ███░░░░░░░░░░░░░░░░░░░░░░ 13.40 %
Sonnet 744 lines █░░░░░░░░░░░░░░░░░░░░░░░░ 03.69 %
Claude-Code 215 lines ░░░░░░░░░░░░░░░░░░░░░░░░░ 01.07 %
Haiku 0 lines ░░░░░░░░░░░░░░░░░░░░░░░░░ 00.00 %
🔎 AI Coding Insights:
🤖 AI-Driven — 99.87% of written lines came from AI
📚 Verbose Prompter — average 16,459 characters per prompt
🔁 Iterative Prompter — average 3 prompts per session
🚀 High AI Trust — 0.16% of changed lines were hand-edited
I Mostly Code in Python
Python 26 repos ██████░░░░░░░░░░░░░░░░░░░ 25.74 %
JavaScript 16 repos ████░░░░░░░░░░░░░░░░░░░░░ 15.84 %
Shell 9 repos ██░░░░░░░░░░░░░░░░░░░░░░░ 08.91 %
TypeScript 7 repos ██░░░░░░░░░░░░░░░░░░░░░░░ 06.93 %
Mermaid 7 repos ██░░░░░░░░░░░░░░░░░░░░░░░ 06.93 %
Last Updated on 23/10/2026 04:38:34 UTC
121.4 million lines across 6,158 hours is 329 a minute without pausing. I have elected not to investigate.